Chris Feather (England)

Original Fairy problems
October’2017 – March’2018

Definitions: (click to show/hide)

No.1272 Chris Feather

original – 04.02.2018

Solutions: (click to show/hide)

white Kh5 Rb6 DGh1 Ph4f5 black Ka3 Pa2a4

ser-h#21*                                   (5+3)
Alphabetic Chess
Double Grasshopper h1

Notify of

Newest Most Voted
Inline Feedbacks
View all comments
Nikola Predrag
Nikola Predrag
February 5, 2018 19:35

Popeye’s apparent “opinion” is irrelevant, since computers can’t think in the first place.
The rules must be unambiguously given BEFORE solving unless it’s all a joke.
Also, the Popeye’s programmers should explain the hidden logic of “null-move”.
For instance, on these examples (Double Grasshopper/Locust):

White DGa8
Black Pa7
Stipulation ~1
White La8
Black Pa7
Stipulation ~1
Condition AntiCirce Circe

Kjell Widlert
Kjell Widlert
February 6, 2018 01:34
Reply to  Nikola Predrag

I have always believed that Popeye prohibits all nul moves, regardless of genre: some people have advocated that nul moves should always be disallowed as they are somehow against the spirit or principle of chess. (I don’t agree with this view, and I have composed nul-move problems myself.)

I have also come to believe that Popeye implements its no-nul-moves rule by always comparing the position on the board before and after every move. But these examples prove that this is not true:
* Popeye says the first position (with DGa8) is a stalemate; White has no move (as DGa8-a6-a8 is prohibited). This is as I expected.
* Popeye says about the second position (with locust a8) that White has the move Hxa7-a6(Ha8)(Pa7), although the position after the move is exactly the same as before.

So does Py only compare before-and-after in fairy forms where nul moves are expected, or does Py not have a general no-nul-moves rule anymore but instead implements specific rules for the DG and some other fairy forms, or is the logic something else?

Over to the Popeye people!

Luce Sebastien
Luce Sebastien
February 5, 2018 20:16

A mistake in the text !
It is C+ with Winchloe if you add the condition
“nul moves prohibited”.

Stephen Emmerson
Stephen Emmerson
February 7, 2018 21:02
Reply to  Luce Sebastien

I’d be interested in understanding what WinChloe is doing here. With all this talk of null moves, I confess I do not see any reason why it makes any difference to this problem. Speaking to Chris, he could offer no reason either.

Can anyone say why this problem requires null moves to be prohibited?

Stephen Emmerson
Stephen Emmerson
February 8, 2018 00:34

Ah, I see…it’s at move 9, by Alphabet rule moves 9 onwards would all be DGa1-a1.

Nikola Predrag
Nikola Predrag
February 5, 2018 23:51

Why this condition (“nul moves prohibited”) is not mentioned under the diagram?
Does the “nul move” have a unique meaning in the first place?

Kjell Widlert
Kjell Widlert
February 6, 2018 01:23
Reply to  Nikola Predrag

The reason the nul-moves-prohibited condition is not mentioned under the diagram is certainly that the common definition of the double grasshopper includes the rule that it may not go back to the same square it started from, producing a nul move (see the Märchenschachlexikon on the web site of Die Schwalbe). But given that WinChloe does not implement this rule, unless you explicitly specify it when entering the problem, it would have been wise to mention it below the diagram.

The double grasshopper was invented by W B Trumper and first published in Feenschach 1968. The text – which seems to be by the editor, not by the inventor – does not mention anything about nul moves.

See also my answer to the earlier question about Popeye.

February 6, 2018 03:33

Kjell, I believe the before & after checks go by the primary part of the move (Lxa7-a6 is not a null move), regardless of the subsequent, secondary part (the fairy effects, here circe rebirths).

Luce Sebastien
Luce Sebastien
February 6, 2018 13:31

I precise that the “logic” of Winchloe is generally to “open” at the maximum the possibilities of a fairy condition, and to propose after to add if you like restrictions.

I think it is a good process !

In some cases the “null move” may be exploited by the composer as a “waiting move”.

Nikola Predrag
Nikola Predrag
February 6, 2018 20:52

Thanks Kjell,
so it is property of the piece and this should be clearly stated in any definition of DG.
Winchloe should respect it (and possibly offer a similar piece, e.g. “null-DG”). But the intrinsic property of a certain type of piece doesn’t apply to all pieces (while a condition would apply).

I still don’t get a clear meaning of the “null move”, it’s even less clear now.

February 6, 2018 21:20

Nikola, I think a null move for a *piece* is when its starting and landing squares coincide (DGa8-a6-a8, or Ra1-a1 on a cylindrical board).

Thing is, fairy problems may have (a lot of) additional fairy conditions stipulated, which may then relocate this piece (and possibly others along with it) further. So in a fairy problem we could talk of a complex move — the primary piece’s move, plus all the secondary fairy side-effects.

I believe Popeye doesn’t support null *primary* moves (of the piece itself). What happens in the second part of the problem’s solution’s *complex* move (the fairy joyride) — only knows…

(Just my understanding. I can see some logic in it. Of course, it would be nice to also have the option “null moves supported” in Popeye.)

February 6, 2018 21:23
Reply to  Adrian

[only knows = only ‘deity’ knows; it was eaten up by the Comments feature, I assume because of the angle brackets surrounding the deity thing in my original text!?…]

Nikola Predrag
Nikola Predrag
February 7, 2018 08:07
Reply to  Adrian

We may imagine examples with various fairy pieces and conditions and test them by Popeye, then try to guess what more or less convincing logic might have caused the obtained results.
But the rules should be clear to a SOLVER without guessing what might be Popeye’s “opinion”. A rough illustration for solving:

White DGh8 Ph7 Pa6 Kc2 Bh2
Black Ka8 Pb7 DGh1
Stipulation H#2
Condition AntiCirce Circe

February 7, 2018 08:55
Reply to  Nikola Predrag

(Vlaicu Crișan, of course, had a lot of fun doing precisely this in his interesting series of Popeye-vs.-WinChloe “feenschach” articles.)

Stephen Emmerson
Stephen Emmerson
February 7, 2018 21:00
Reply to  Nikola Predrag

“then try to guess what more or less convincing logic might have caused the obtained results”…or, one may examine the source code, since it is freely available and accessible! 🙂

I understand that’s not ideal, and it would be better if behaviour could be documented (remembering that Popeye has evolved over three decades, and some of its behaviour has arisen by accidental combination of fairy genres that weren’t envisaged by the programmers of each individual part). The best situation of all would be that all variations of rules were allowed through Popeye’s input language. No solving program though can anticipate everything, unless it sacrifices simplicity (of course, one can always write a program or modify Popeye’s source code to test something specific).

As far as I can see there is no general prohibition of null moves in Popeye, but the implementation of DG will ensure no null moves will be generated. There is this comment in the source code:

/* W.B.Trumper feenschach 1968 – but null moves will not be allowed by Popeye */

which suggests it’s deliberate that null DG moves are forbidden, though speaking to Chris Feather he is under the impression that W.B.Trumper never intended them, either. But I think the DG may have been invented more than once…

In short, *I think* the general comments about null moves in Popeye are a red herring. Popeye generally allows null moves. But, it does not allow a DG (and some other pieces) to move to its departure square (which might be by design), and it does not offer the facility to allow these if required for a particular problem.

Features can be requested at Popeye’s website. Of course, since it is practically only Thomas Maeder maintaining it now, one may have to be patient, or get involved!

Nikola Predrag
Nikola Predrag
February 8, 2018 03:18

my question is not about a computer program.
It’s about the publication of a problem – it must clearly present the rules to a solver. The definition of DG (above the diagram) says nothing about null-moves.
One must read the solution and comment to hear about it, but without a clear definition.
(And what around the diagram indicates the rules “by Popeye” and not “by Winchloe”?)

Wishing to understand the comment, one may put wDGh8,bDGh1,wPh7,bPh2 and try h~1, Anticirce+Circe. Just hoping for some help from a machine 🙂

Stephen Emmerson
Stephen Emmerson
February 8, 2018 23:07
Reply to  Nikola Predrag

OK, I understand more. I agree entirely that the diagram should make it clear what the rules are. Saying “Popeye rules” or “WinChloe rules with No Null moves” only helps if those programs clearly define their rules. Presumably “DG” on its own must be unclear as to whether null moves are allowed, or else the two programs would not have ended up doing things differently in the first place.

In the absence of a standard reference source for fairy definitions, perhaps Julia could amend the particular DG definition above the diagram to clarify “In this problem, the DoubleGrasshopper may not make a null move”?

As for Kjell’s comment that he thinks Popeye forbids null moves in general, that is not the case. I think it would be hard to write the code to optionally forbid that.

It does disallow in the particular case of DG moves. That is much easier to (optionally) allow, I might attempt the code change myself.

Vlaicu Crisan
Vlaicu Crisan
February 9, 2018 18:23

My guess is Julia uses in general the definitions listed on Christian Poisson’s Problemesis website.

As rightly stated by Kjell, the DG definition from Die Schwalbe includes the following remark: The hurdle can be the same piece in both G jumps, but at the end of the move the DG can not land on its starting field (i.e. null moves are not allowed).

Using the definition of fairy pieces/conditions from Problemesis, this problem has no solution. In this case, it would be necessary to alter the DG definition listed above the diagram.

For practical reasons, I don’t think the composers should be always obliged to specify the definition(s) of the fairy pieces / conditions used in their compositions. As a matter of common sense, it should be enough to specify the computer program (and version) they used to check their composition. Although we all know these software applications are not quite bug-free…

For the readers aware about the feenschach articles mentioned by Adrian, I will prepare another episode. The series started with a solving contest featuring 6 innocent looking #1 being C+ both in Popeye and Winchloe, but the solution listed by Popeye was entirely different from the solution listed by Winchloe. The solvers were asked to discover the correct solution and explain the reason.

Back to 1272: I liked a lot the way the set indirect mate becomes a direct pin-mate! Is it possible to show the same contents without using ABC?

Geoff Foster
Geoff Foster
February 9, 2018 23:01
Reply to  Vlaicu Crisan

Without ABC it would be difficult to prevent the set mate from returning e.g. 1.a1=S 2.Sc2 3.Sb4 4.Sa2 Rb7#.

Juraj Lörinc
Juraj Lörinc
February 10, 2018 01:54
Reply to  Geoff Foster

When hoppers are involved (as here), one of irreversible moves usable as B1 is jumping away with blocking hopper. Many examples in existence, let me show one where CJF was involved (with me):

4th HM StrateGems 2005
White : Kc5 Sb1 KAa5
Black : Ka1 KAa6a2b2
1+3 kangaroos

1.KAa7 2.Ka2 3.KAa1 4.Kb3 5.Kc2 6.Kd3 7.Ke3 8.KAf2 9.Ke2 10.KAg2 11.KAd2 12.Ke1 13.KAf1 Sc3#

February 6, 2018 22:25

hm… Does Chris sends by snailmail?

S. K. Balasubramanian
S. K. Balasubramanian
March 12, 2018 14:19

I wish to quote one of the problems of mine with Vlaicu Crisan published in The Problemist (Prob. F-3033,March 2013):

Wh: Kf8, Rh1, Bf1, Sd4, Pb2
Bl: Kh3, Rh2, Bg2, Pf2, Ph5
H#2, (Circe+T&M+KobulKings) 2 Sol (5+5)

1.B*f1-g2(+wBf1[wK=rB] rBg7 2.R*h1-h2(+wRh1) [wK=rR] B*g2-f1(+bBc8)[bK=rB]#
1.R*h1-h2(+wRh1)[wK=rR] rRf4 2.B*f1-g2(+wBf1)[wk=rB] R*h2-h1(+rRh8)[bK=rR]#

Popeye solves it though B1 & B2 in both the solutions look like null moves. Here the position just after and just before B1 & B2 are the same. Ofcourse, the wK changes its power, so they may not be called exactly null moves. But, I think most probably if the kobul king condition is removed and only circe & T&M exist, and if there exists a solution with Bg2-bf1-g2 (+wBg1), this null move will be indicated by the Popeye.

Would love your thoughts, please comment.x